Ask Your Question
1

The screen shows 'Cannot get disk parameters' while deploying Mirantis OpenStack using Fuel

asked 2014-10-15 04:47:29 -0500

Wayne Yeh gravatar image

updated 2014-10-16 21:55:02 -0500

I have deployed a master node of Fuel 5.1, and want to deploy a runnable OpenStack environment using it. I've got two servers with the same hardware (Intel Xeon E5-2620 x2, 32GB RAM, 4TB HDD, 2 NICs). On the web UI of Fuel, I create an OpenStack environment with adding the two servers as its nodes. One node is deployed with Controllder + Storage - Cinder LVM + Telemetry - MongoDB, and the other is Compute + Storage - Ceph OSD. However, the progress of deployment never ends. One of the node is deployed successfully, but the other is not. I have tried several times with switching roles between the two nodes, but no luck. The fine node always success, and the broken node can never be deployed correctly.

Following is the messages showed on the screen of the broken node. The messages are showed after ubuntu is installed automatically through Fuel, and the deployment progress of this node is stuck at 29% on Fuel's web UI.

BOOT SERVER IP: 172.29.0.1

PXELINUX 4.02 2010-07-21  Copyright (C) 1994-2010 H. Peter Anvin et al
!PXE entry point found (we hope) at 8E30:0106 via plan A
UNDI code segment at 8E30 len 4810
UNDI data segment at 86EE len 7420
Getting cached packet  01 02 03
My IP address seems to be AC1D0003 172.29.0.3
ip=172.29.0.3:172.29.0.1:172.29.0.1:255.255.0.0
BOOTIF=01-bc-5f-f4-ca-4a-59
SYSUUID=.......
TFTP prefix:
Trying to load: pxelinux.cfg/01-bc-5f-f4-ca-4a-59
Cannot get disk parameters
boot: _

I have no idea about the difference between the two servers. Is there anything I can check to get it right?

Many thanks!

Edit:

The two server nodes which are used for deploying OpenStack are named as Untitled (4a:59) and Untitled (49:87) by Fuel. I have created a new OpenStack environment and try to deploy it to the two server nodes again. The OpenStack environment I'm going to deployed is as follows.

OpenStack Release: Icehouse on Ubuntu 12.04.4 (2014.1.1-5.1)
Deployment Mode: Multi-node
Compute: KVM
Network: Neutron with GRE segmentation
Storage Backend: Default (Cinder), Default (Glance)
Additional Services: Install Ceilometer (OpenStack Telemetry)

The roles to the two nodes are listed as follows.

Untitled (49:87): CONTROLLER‧CINDER‧MONGO
Untitled (4a:59): COMPUTE‧CEPH-OSD

Before deploying changes, the logs of the two nodes are as follows. (all sources except messages show nothing in the error log)

image description

image description

After clicking Deploy Changes, the result of deployment is as follows.

image description

Node Untitled (4a:59) shows the same message as the above description. It is stuck in the reboot phase after ubuntu is installed, but before installing OpenStack. No additional error log is present in both Bootstrap logs and Other install logs source categories.

For logs in Fuel Master, there is only one error log in Web backend (shows as follows), which I don't think it is relevant, and no error log in other sources.

    [7fa17a9d0740] (manager ...
(more)
edit retag flag offensive close merge delete

Comments

Are you using Software based Raid? Or do you have a raid controller that is actually "emulating" hardware raid?

mpetason gravatar imagempetason ( 2014-10-15 11:16:27 -0500 )edit

@mpetason The server node does have RAID controller. I'm not quite sure it is a software RAID or a hardware RAID though I was told that it has a RAID card. However, I have disabled all of the RAID configurations to both two nodes before I start deploying OpenStack using Fuel.

Wayne Yeh gravatar imageWayne Yeh ( 2014-10-15 22:16:29 -0500 )edit

@Wayne Unfortunately, with virtually no errors, seems rather a hardware problem. We assume that disks and raid of 4a:59 work properly and a bare Ubuntu ISO can be installed on it, correct? Have you ever tried it?

Fsoppelsa gravatar imageFsoppelsa ( 2014-10-16 10:21:08 -0500 )edit

@Fsoppelsa 4a:59 was installed with ubuntu 12.04 using official ISO image, and worked fine before I reset it for deploying OpenStack. I also tried exchanging hard disks between 4a:59 and 49:89, but situation still remains.

Wayne Yeh gravatar imageWayne Yeh ( 2014-10-16 10:59:36 -0500 )edit

How did you partition your disks? In the Fuel UI, when you select a node, you can Configure Disks. Is this disks configuration the same for both nodes? Where did you install Base System for both? Could you please paste a screenshot?

Fsoppelsa gravatar imageFsoppelsa ( 2014-10-16 12:03:42 -0500 )edit

4 answers

Sort by » oldest newest most voted
1

answered 2014-10-15 08:52:38 -0500

Hi Wayne,

which one is the node that completes, and which doesn't?

The second node stucks at 29% while still installing the operating system, or while installing OpenStack?

Your roles configuration is correct. Hardware is fine to do the thing, assuming it doesn't have problems itself.

Can you find relevant information into the Fuel UI panel Logs for the Fuel Master at ERROR level for each of the sources?

So, to help us to provide an answer, please look into logs, check if on the master all the services are running fine (output of dockerctl check) and paste some relevant information about the OpenStack environment you're going to deploy:

  • Ubuntu or CentOS?
  • HA or Multinode?
  • Glance on default or on Ceph?
  • Networking choice.
  • Other add ons (like Murano or Sahara)?

Fabrizio -Mirantis Fuel team

edit flag offensive delete link more

Comments

Thanks for your reply. I have updated more details about my deployment environment and logs in the post.

Wayne Yeh gravatar imageWayne Yeh ( 2014-10-15 23:23:50 -0500 )edit

This doesn't look at all like an answer to me. Although the hint to run dockerctl check may be useful.

tudor gravatar imagetudor ( 2015-01-29 17:53:54 -0500 )edit

Got this problem too on a DL360p gen8 Server. I've configured the base system to install on a SD card. Still not able to resolve the problem through this thread.

iGene gravatar imageiGene ( 2016-01-29 23:42:04 -0500 )edit
2

answered 2014-12-02 19:16:58 -0500

lfz gravatar image

I had the same error verbatim. In my case it was because I had disabled all HDD boot options in the BIOS with only PXE boot. When I left PXE as first boot and made the OS disk second boot, problem solved.

edit flag offensive delete link more
0

answered 2015-08-27 10:59:37 -0500

i have the same issue but i am using SW RAID will it be the problem then?

edit flag offensive delete link more
0

answered 2016-06-03 13:14:49 -0500

fbirlik gravatar image

lfz's answer is correct.

I assume that intention with PXE boot is just for bootstrap of unconfigured nodes. After OS installation is complete at target node, PXE configuration fails to boot the device with bootstrap in order to let it fall into installed OS on the hard drive. In case hard disk is not a valid boot option and PXE fails, system is stuck at boot.

I couldn't find exact doc/code for hypervisor installation, but bare metal deployment is similar. First PXE bootstrap for installation, second boot is from hard drive (http://docs.openstack.org/developer/ironic/deploy/user-guide.html#understanding-bare-metal-deployment (http://docs.openstack.org/developer/i...))

Your BIOS boot order should be first PXE and hard drive boot option should stay enabled.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2014-10-15 04:47:29 -0500

Seen: 3,934 times

Last updated: Oct 16 '14